arXiv subjects

Yebo Feng

Publications and source records attributed to Yebo Feng.

At least 19 recordsLinked to original sources

arXiv2026

A Visionary Look at Vibe Researching

Vibe researching is an emerging paradigm in which human researchers provide high-level direction and critical judgment while LLM-based agents handle the labor-intensive execution of literature review, experimentation, data analysis, and manuscript drafting. Inspired by the "vibe coding" movement in software engineering, it occupies a middle ground between traditional manual research and fully autonomous AI research systems. This paper defines the concept, describes its methodology (multi-agent architectures, memory, tool use, retrieval-augmented generation, and the human's role as orchestrator), identifies seven technical limitations, weighs its positive and negative societal impacts, and maps each problem to a concrete future direction. Our goal is to provide the research community with a clear and honest map of the territory so that the conversation about responsible adoption can start from shared ground.

cs.CY↗

arXiv2026

Sealing the Audit-Runtime Gap for LLM Skills

Large language model (LLM) ecosystems such as Claude Code and ChatGPT increasingly rely on skills: packages of natural-language instructions and executable tools. Once in the LLM's context, skill content cannot be reliably separated from trusted instructions, and a skill's executable side can invoke privileged actions, exposing the skill supply chain to injection, tampering, and rug-pull attacks. Existing defenses are stage-bound: centralized signing, audit reports unbound from the runtime artifact, or policy engines that cannot attest to what was approved. We present SIGIL, the first framework that seals the audit-runtime gap for LLM skills. SIGIL delivers verifiable hosting through a tamper-evident, decentralized on-chain registry from which LLMs fetch skills directly. The registry admits four publication types, Transparent, Licensed, Sealed, and Committed, spanning plaintext public distribution, monetized access, custodial use, and off-chain workflows; before admission, every skill is vetted by a Decentralized Autonomous Organization (DAO) audit committee that supports pluggable auditing methods under a stake-and-slash economic model. At load time, SIGIL delivers verified loading through a skill verification protocol executed by a Skill Verification Loader (SVL) embedded as the mandatory loading path: the SVL retrieves and decrypts the skill as its type requires, verifies its integrity against the on-chain record, and enforces its permission manifest before context injection. We evaluate SIGIL on a real-world deployment against 1,023 in-the-wild skills spanning six attack types. At load time, the SVL verifies each skill's integrity against its on-chain record and enforces its approved permission manifest, completing batched verification under 86 ms. Together, these results show that LLM skills can be cryptographically bound from publication through runtime at practical cost.

cs.CR↗

arXiv2026

The Anonymity Gap: Understanding Real Privacy in Shielded UTXO-based Protocols for DeFi

Shielded UTXO-based protocols are becoming a core form of privacy infrastructure for DeFi. Unlike mixers that organize privacy mainly around deposits and withdrawals, these protocols allow assets, once inside the shielded pool, to continue moving and being re-spent within the hidden state, and to become public only when users withdraw or interact with public DeFi protocols. Their anonymity is therefore no longer a flat pool-size problem, but a provenance problem that propagates across the note/UTXO, proof, and transaction layers. Yet, a unified analysis framework for this setting is still missing. We propose a layered system model and an analysis pipeline that uses prior history as the temporal baseline, applies cumulative pruning and cross-proof propagation to each proof's Commitment Set, and recursively traces the survivors through historical hidden-state transitions to derive the final transaction-level Anonymity Set Size. We evaluate our methodology on the complete on-chain histories of all four Railgun production deployments and five independent Hinkal pools across six EVM chains, analyzing 186,356 unshielding spend transactions. Using only public protocol traces and constraints, our non-heuristic analysis yields mean Anonymity Set Size reductions of 40.1%-59.0% relative to each deployment's temporal baseline; 3,679 transactions retain at most 10 addresses, including 1,228 singletons. Public token constraints are the strongest and most stable source of pruning in both protocols, while the effects of tree number, proof roots, and value constraints vary with protocol design and historical state. Together with representative cases, these results reveal interpretable anonymity-loss patterns and implications for user behavior and future protocol design.

cs.CR↗

arXiv2026

LLM-Powered Multi-Agent System for Automated Crypto Portfolio Management

Cryptocurrency portfolio management requires the fusion of heterogeneous multi-modal signals, including structured price and on-chain time series, unstructured news text, and technical indicators, under high-volatility and real-time constraints. While deep learning approaches show predictive capability, their opacity limits practical adoption, and single large language model (LLM) agents struggle to process the breadth of modality-specific inputs needed for robust decision-making. We propose a multi-agent system (MAS) framework in which three modality-specialised agents, a Crypto Agent for market dynamics, a News Agent for weekly news sentiment, and a Trading Agent for signal fusion and portfolio execution, decompose the task across three communication architectures: hierarchical, collaborative, and debate. We evaluate four capability configurations: zero-shot, chain-of-thought (CoT), retrieval-augmented generation (RAG), and skill-augmented. In a 52-week backtest over calendar year 2025 across the top 15 L1 blockchain native cryptocurrencies by market capitalisation as of January 2025, the best configuration, Hierarchical (Skill), achieves a cumulative return of 133.52% and a Sharpe ratio of 1.502, outperforming single-agent variants, passive benchmarks, and deep learning baselines. An ablation study identifies the Crypto Agent as the most critical component, with its removal reducing cumulative return by 42.57 percentage points. A cross-model comparison further shows that MAS outperforms the single-agent baseline under GPT-4o, GPT-5, and Claude Sonnet 4.5, suggesting that the benefit of multi-agent coordination is model-agnostic. Unlike black-box deep learning models, every portfolio decision is traceable to explicit agent reasoning, offering an interpretable and effective approach to multi-modal cryptocurrency portfolio management.

q-fin.TR↗

arXiv2026

Seeing Is Not Screening: Multimodal Hidden Instruction Attacks on Agent Skill Scanners

Agent skills are emerging as an important attack surface in LLM-based systems. Through an empirical study of existing skill scanners, we find that current defenses primarily rely on textual descriptions, manifests, and source code as the main signals for security analysis, which can leave visually conveyed malicious intent insufficiently examined. This creates a practical blind spot: harmful operational instructions hidden in images may bypass scanning while still being recoverable by multimodal agents during deployment. To systematically investigate this threat, we propose SkillCamo, a document-mediated multimodal instruction attack that conceals malicious instructions within images bundled with a skill while rewriting the surrounding documentation to naturally reference those images as part of the normal workflow. Thus, the attack does not rely on the image alone, but on the joint interpretation of textual guidance and visual payload at execution time. To defend against such attacks, we further propose ExecScan, an execution-grounded multimodal scanning module that performs intent extraction, behavior reconstruction, abuse assessment, and deliberative execution simulation over skill artifacts. ExecScan jointly analyzes documentation, code, referenced resources, and visual content to recover hidden instructions, reconstruct executable behavior chains, and identify downstream risks such as exfiltration, destruction, persistence, deception, and privilege escalation. Extensive experiments show that image-hidden malicious instructions challenge existing skill scanners, while ExecScan can improve the skill scanning performance.

cs.CR↗

arXiv2026

Execution-Grounded Security Testing for Coding Agents in Software Engineering Pipelines

Coding agents are increasingly integrated into system operations, where their tool use can directly modify project artifacts, execution environments, and the underlying system. For example, if a coding agent inserts a hook into a system startup or configuration script, that change can persist after the interaction, be triggered later, and abuse delegated user or system privileges to modify the system. This makes security testing a system problem: the key question is not only what the agent says, but what it actually does to the surrounding environment. We present an execution-grounded red-team testing framework for probing this execution-layer security boundary using observable sandbox evidence, including tool invocations, runtime traces, and file-system diffs. Our framework embeds target unsafe operations into routine software engineering workloads, including unit testing, regression testing, crash reproduction, and validation, and uses an execution oracle to guide refinement when an initial probe is rejected or fails. Across multiple agent frameworks and model backbones, our red-team workload reformulation substantially increases verified unsafe execution, reaching 73.61% on code carriers and 53.93% on text carriers. These results show that coding agents in system operations remain insecure under task disguise: once risky intent is hidden inside plausible engineering tasks, the agent can be induced to carry out unsafe actions on the surrounding system. More broadly, coding agents in system operations still demand stronger security testing and safeguards.

cs.AI↗

arXiv2026

Babel: Jailbreaking Safety Attention via Obfuscation Distribution Optimized Sampling

Despite rigorous safety alignment, Large Language Models (LLMs) remain vulnerable to jailbreak attacks. Existing black-box methods often rely on heuristic templates or exhaustive trials, lacking mechanistic interpretability and query efficiency. In this study, we investigate an intrinsic vulnerability in the safety mechanisms of LLMs, where safety alignment relies on a small set of sparsely distributed attention heads, leaving much of the representational space weakly monitored. We formalize this phenomenon with a mathematical jailbreaking model that characterizes the delicate boundary of effective text obfuscation and analytically explains observed jailbreak behaviors. Guided by this model, we propose Babel, an efficient black-box attack framework that exploits the identified safety gap through systematic obfuscation sampling with iterative, feedback-driven distribution refinement, enabling reliable and high-success jailbreak attacks without access to model internals. Comprehensive evaluations on frontier commercial models demonstrate that Babel achieves state-of-the-art attack success rates and superior query efficiency. Specifically, compared to state-of-the-art methods, Babel increases the attack success rate on GPT-4o from 41.33% to 82.67% and on Claude-3-5-haiku from 38.33% to 78.33% within an average of 40 queries, providing a robust red-teaming methodology for LLMs safety research.

cs.CR↗

arXiv2026

Don't Trust Your Upstream: Exploiting LLM Multi-Agent System via Topology-Guided Adversarial Propagation

The digital world is witnessing the rapid rise of LLM-based multi-agent systems (MASs) and their powerful applications. However, their security remains insufficiently understood, as existing evaluations are largely limited to narrow attack settings and may substantially underestimate the real risks of MAS deployments. Inspired by the MAS inter-agent dependencies, where upstream outputs are reinterpreted and executed by downstream agents, we propose a topology-aware attack scheme that propagates adversarial contamination from exposed edge agents to high-privilege agents to induce malicious behaviors. By combining topology reconnaissance, contamination propagation modeling, and hierarchical payload encapsulation, our approach overcomes the key challenges of black-box attacks and makes such multi-hop compromise practical. Experiments show that our approach achieves success rates of 40\%--78\% on three widely-used MAS frameworks under five topologies, and 85\% on two real-world MAS applications across 20 representative scenarios. The results reveal fundamental vulnerabilities in MASs that have been overlooked by prior studies. Based on these findings, we propose a topology-trust mitigation that blocks 94.8\% of such composite attacks.

cs.CR↗

arXiv2026

EvoPoC: Automated Exploit Synthesis for DeFi Smart Contracts via Hierarchical Knowledge Graphs

Smart contract vulnerabilities in Decentralized Finance caused over billions of dollars losses every year, yet the security community faces a critical bottleneck: identifying a vulnerability is not the same as proving it is exploitable. Manual PoC construction is prohibitively labor-intensive, leaving most disclosed vulnerabilities unverified and protocols exposed long before mitigation is applied. In this paper, we propose \sys, a knowledge-driven agentic system for end-to-end contract vulnerability detection and exploit synthesis. Our core insight is that exploit synthesis is not a code generation task but a \emph{structured reasoning problem} that requires grounded knowledge of protocol semantics, failure root cause, and exploit primitives. \sys organizes this knowledge into a \emph{Hierarchical Knowledge Graph} (HKG) that serves as structured memory for LLM-guided multi-hop reasoning. To validate exploit feasibility beyond code synthesis, \sys employs a two-stage validation framework that checks exploit-path reachability via SMT solving and profit realizability via asset-level state simulation, ensuring generated PoCs satisfy both logical and economic viability constraints. Evaluated on 88 real-world DeFi attacks and 72 audited projects (2,573 contracts), \sys achieves 98\% recall and 0.9 F1-score in detection, and a 96.6\% exploit success rate (ESR), reproducing 85 historical exploits and recovering over \$116.2M revenue. \sys outperforms SOTA fuzzers (\textsc{Verite}, \textsc{ItyFuzz}) by up to $5\times$ in ESR and $300\times$ in recoverable value, and the LLM-based exploit generator \textsc{A1} by $2\times$ and $8.5\times$ respectively. In bug bounty evaluation, \sys identified 16 confirmed 0-day vulnerabilities, helping secure over \$70.6M and earning \$2,900 in bounties.

cs.CR↗

arXiv2026

Systematizing Blockchain Research Themes and Design Patterns: Insights from the University Blockchain Research Initiative (UBRI)

The rapid expansion of blockchain and digital asset ecosystems has intensified the challenge of translating academic research into deployable systems and regulatory frameworks. While advances in cryptography, consensus, digital assets, and governance are substantial, institutional mechanisms that sustain research-to-deployment translation at ecosystem scale remain comparatively under-theorized. This paper examines the architectural and coordination patterns that enable such translation, using the University Blockchain Research Initiative (UBRI) network as a representative case of long-term academic and industry collaboration. Drawing on research outputs and convenings from 2022 to 2025, we synthesize recurring design tensions across technical and institutional domains, including scalability versus security, decentralization versus governance, and privacy versus compliance. Rather than cataloging individual projects, we abstract system-level themes that connect research contributions to deployment constraints and policy adaptation, providing a structured lens for understanding how academic research informs production architectures, regulatory development, and ecosystem resilience in emerging decentralized infrastructures.

cs.DC↗

arXiv2026

Resisting Manipulative Bots in Meme Coin Copy Trading: A Multi-Agent Approach with Chain-of-Thought Reasoning

Copy trading has become the dominant entry strategy in meme coin markets. However, due to the market's extremely illiquid and volatile nature, the strategy exposes an exploitable attack surface: adversaries deploy manipulative bots to front-run trades, conceal positions, and fabricate sentiment, systematically extracting value from naïve copiers at scale. Despite its prevalence, bot-driven manipulation remains largely unexplored, and no robust defensive framework exists. We propose a manipulation-resistant copy-trading system based on a multi-agent architecture powered by a multi-modal large language model (LLM) and chain-of-thought (CoT) reasoning. Our approach outperforms zero-shot and most statistic-driven baselines in prediction accuracy as well as all baselines in economic performance, achieving an average copier return of 3% per meme coin investment under realistic market frictions. Overall, our results demonstrate the effectiveness of agent-based defenses and predictability of trader profitability in adversarial meme coin markets, providing a practical foundation for robust copy trading.

cs.AI↗

arXiv2026

Can Distillation Mitigate Backdoor Attacks in Pre-trained Encoders?

Self-Supervised Learning (SSL) has become a prominent paradigm for pre-training encoders to learning general-purpose representations from unlabeled data and releasing them on third-party platforms for broad downstream deep learning tasks. However, SSL is vulnerable to backdoor attacks, where an adversary may train and distribute poisoned pre-training encoders to contaminate the downstream models. In this paper, we study a defense mechanism based on distillation against poisoned encoders in SSL. Traditionally, distillation transfers knowledge from a pre-trained teacher model to a student model, enabling the student to replicate or refine the teacher's learned representations. We repurpose distillation to extract benign knowledge and remove backdoors from a poisoned pre-trained encoder to produce a clean and reliable pre-trained model. We conduct extensive experiments to evaluate the effectiveness of distillation in mitigating backdoor attacks on pre-trained encoders. Based on two state-of-the-art backdoor attacks and four widely adopted image classification datasets, our results demonstrate that distillation reduces the attack success rate from 80.87% to 27.51%, with only a 6.35% drop in model accuracy. Furthermore, by comparing four teacher architectures, three student models, and six loss functions, we find that the distillation with fine-tuned teacher networks, warm-up-based student training, and attention-based distillation losses yield the best performance.

cs.LG↗

arXiv2026

Enshrined Proposer Builder Separation in the presence of Maximal Extractable Value

In blockchain systems operating under the Proof-of-Stake (PoS) consensus mechanism, fairness in transaction processing is essential to preserving decentralization and maintaining user trust. However, with the emergence of Maximal Extractable Value (MEV), concerns about economic centralization and content manipulation have intensified. To address these vulnerabilities, the Ethereum community has introduced Proposer Builder Separation (PBS), which separates block construction from block proposal. Later, enshrined Proposer Builder Separation (ePBS) was also proposed in EIP-7732, which embeds PBS directly into the Ethereum consensus layer. Our work identifies key limitations of ePBS by developing a formal framework that combines mathematical analysis and agent-based simulations to evaluate its auction-based block-building mechanism, with particular emphasis on MEV dynamics. Our results reveal that, although ePBS redistributes responsibilities between builders and proposers, it significantly amplifies profit and content centralization: the Gini coefficient for profits rises from 0.1749 under standard PoS without ePBS to 0.8358 under ePBS. This sharp increase indicates that a small number of efficient builders capture most value via MEV-driven auctions. Moreover, 95.4% of the block value is rewarded to proposers in ePBS, revealing a strong economic bias despite their limited role in block assembly. These findings highlight that ePBS exacerbates incentives for builders to adopt aggressive MEV strategies, suggesting the need for future research into mechanism designs that better balance decentralization, fairness, and MEV mitigation.

cs.DC↗

arXiv2025

SoK: Design, Vulnerabilities, and Security Measures of Cryptocurrency Wallets

With the advent of decentralised digital currencies powered by blockchain technology, a new era of peer-to-peer transactions has commenced. The rapid growth of the cryptocurrency economy has led to increased use of transaction-enabling wallets, making them a focal point for security risks. As the frequency of wallet-related incidents rises, there is a critical need for a systematic approach to measure and evaluate these attacks, drawing lessons from past incidents to enhance wallet security. In response, we introduce a multi-dimensional design taxonomy for existing and novel wallets with various design decisions. We classify existing industry wallets based on this taxonomy, identify previously occurring vulnerabilities and discuss the security implications of design decisions. We also systematise threats to the wallet mechanism and analyse the adversary's goals, capabilities and required knowledge. We present a multi-layered attack framework and investigate 84 incidents between 2012 and 2024, accounting for $5.4B. Following this, we classify defence implementations for these attacks on the precautionary and remedial axes. We map the mechanism and design decisions to vulnerabilities, attacks, and possible defence methods to discuss various insights.

cs.CR↗

arXiv2025

AMM-based DEX on the XRP Ledger

Automated Market Maker (AMM)-based Decentralized Exchanges (DEXs) are crucial in Decentralized Finance (DeFi), but Ethereum implementations suffer from high transaction costs and price synchronization challenges. To address these limitations, we compare the XRP Ledger (XRPL)-AMM-Decentralized Exchange (DEX), a protocol-level implementation, against a Generic AMM-based DEX (G-AMM-DEX) on Ethereum, akin to Uniswap's V2 AMM implementation, through agent-based simulations using real market data and multiple volatility scenarios generated via Geometric Brownian Motion (GBM). Results demonstrate that the XRPL-AMM-DEX achieves superior price synchronization, reduced slippage, and improved returns due to XRPL's lower fees and shorter block times, with benefits amplifying during market volatility. The integrated Continuous Auction Mechanism (CAM) further mitigates impermanent loss by redistributing arbitrage value to Liquidity Providers (LPs). To the best of our knowledge, this study represents the first comparative analysis between protocol-level and smart contract AMM-based DEX implementations and the first agent-based simulation validating theoretical auction mechanisms for AMM-based DEXs.

cs.CE↗

arXiv2025

ProvX: Generating Counterfactual-Driven Attack Explanations for Provenance-Based Detection

Provenance graph-based intrusion detection systems are deployed on hosts to defend against increasingly severe Advanced Persistent Threat. Using Graph Neural Networks to detect these threats has become a research focus and has demonstrated exceptional performance. However, the widespread adoption of GNN-based security models is limited by their inherent black-box nature, as they fail to provide security analysts with any verifiable explanations for model predictions or any evidence regarding the model's judgment in relation to real-world attacks. To address this challenge, we propose ProvX, an effective explanation framework for exlaining GNN-based security models on provenance graphs. ProvX introduces counterfactual explanation logic, seeking the minimal structural subset within a graph predicted as malicious that, when perturbed, can subvert the model's original prediction. We innovatively transform the discrete search problem of finding this critical subgraph into a continuous optimization task guided by a dual objective of prediction flipping and distance minimization. Furthermore, a Staged Solidification strategy is incorporated to enhance the precision and stability of the explanations. We conducted extensive evaluations of ProvX on authoritative datasets. The experimental results demonstrate that ProvX can locate critical graph structures that are highly relevant to real-world attacks and achieves an average explanation necessity of 51.59\%, with these metrics outperforming current SOTA explainers. Furthermore, we explore and provide a preliminary validation of a closed-loop Detection-Explanation-Feedback enhancement framework, demonstrating through experiments that the explanation results from ProvX can guide model optimization, effectively enhancing its robustness against adversarial attacks.

cs.CR↗

arXiv2025

Hedge Funds on a Swamp: Analyzing Patterns, Vulnerabilities, and Defense Measures in Blockchain Bridges

Blockchain bridges have become essential infrastructure for enabling interoperability across different blockchain networks, with more than $24B monthly bridge transaction volume. However, their growing adoption has been accompanied by a disproportionate rise in security breaches, making them the single largest source of financial loss in Web3. For cross-chain ecosystems to be robust and sustainable, it is essential to understand and address these vulnerabilities. In this study, we present a comprehensive systematization of blockchain bridge design and security. We define three bridge security priors, formalize the architectural structure of 13 prominent bridges, and identify 23 attack vectors grounded in real-world blockchain exploits. Using this foundation, we evaluate 43 representative attack scenarios and introduce a layered threat model that captures security failures across source chain, off-chain, and destination chain components. Our analysis at the static code and transaction network levels reveals recurring design flaws, particularly in access control, validator trust assumptions, and verification logic, and identifies key patterns in adversarial behavior based on transaction-level traces. To support future development, we propose a decision framework for bridge architecture design, along with defense mechanisms such as layered validation and circuit breakers. This work provides a data-driven foundation for evaluating bridge security and lays the groundwork for standardizing resilient cross-chain infrastructure.

cs.ET↗

arXiv2025

A Vision for Auto Research with LLM Agents

This paper introduces Agent-Based Auto Research, a structured multi-agent framework designed to automate, coordinate, and optimize the full lifecycle of scientific research. Leveraging the capabilities of large language models (LLMs) and modular agent collaboration, the system spans all major research phases, including literature review, ideation, methodology planning, experimentation, paper writing, peer review response, and dissemination. By addressing issues such as fragmented workflows, uneven methodological expertise, and cognitive overload, the framework offers a systematic and scalable approach to scientific inquiry. Preliminary explorations demonstrate the feasibility and potential of Auto Research as a promising paradigm for self-improving, AI-driven research processes.

cs.AI↗